This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: gcc compile-time (multibyte issue)


<<It doesn't count raw input characters; it counts raw input bytes. Counting
characters would get this case correct for UTF-8, and that would get closer
than counting bytes. If you want to count positions, Markus Kuhn's public
domain wcwidth could be dropped in, and would fit how it would be displayed
on most consoles and when printed.
>>

Right it counts raw input bytes. In my judgment it would be an unacceptable
extra burden on the lexer to count logical characters. This is only a style
checking option, and the fact that it does not apply to UTF-8 does not bother
me. If someone wants to check style by counting real characters then that
can be done perfectly well by an external tool. Basically the style checing
options that are in GNAT are those necessary to enforce GNAT's own internal
style. We make them more generally available, since why not, but there is
no attempt to generalize into a full style checker -- that's more appropriate
as a separate tolol anyway.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]